Much has been written about the current unprecedented expansion of new generic Top Level Domains, known as new gTLDs, by ICANN, which will radically change how we use and view the internet. However, most legal publications have tended to focus heavily on legal rights protection, in particular in relation to trade marks, often at the expense of some of the other issues which may also have an impact on the business world. One of these neglected issues relates to the notion of ‘communities’ and the new gTLD program. This article aims to rectify this and clarify some common misconceptions.
The new gTLD program
ICANN was set up in 1998 by the US Government and is the international body responsible for ensuring the stability and security of the Internet. The body within ICANN that develops policies in relation to gTLDs, such as .COM and .NET, is known as the Generic Names Supporting Organisation (GNSO). In August 2007, the GNSO published its final report on the introduction of new gTLDs. For various reasons this advocated the introduction of additional gTLDs (at the time there were only 21) in an orderly, timely and predictable way.
The report set down principles, recommendations and guidelines for the introduction of new gTLDs and provided that ‘the evaluation and selection criteria for new gTLD registries should reflect the principles of fairness, transparency and non-discrimination’. In the event that the same gTLD string was applied for by more than one applicant, the report stated that a claim to support a community by one party would be a reason to award priority to that application. Furthermore, applications would be rejected if there was substantial community opposition to them. It can therefore be seen that, right from the early days, it was envisaged that communities would be given special significance within ICANN’s new gTLD program and protected by a dual mechanism, on the one hand prioritising applications that were made by genuine communities, and on the other weeding out applications that were harmful to such communities.
The ICANN Board adopted the GNSO’s recommendations in 2008. This marked the start of ICANN’s new gTLD program and resulted in the publication of the first Applicant Guidebook in October 2008. The Applicant Guidebook is over 300 pages long and sets out detailed guidelines and procedures for potential new gTLD applicants. It went through ten different versions, all of which were subject to public comment in accordance with ICANN’s bottom-up consensus model (in fact there were 47 public comment periods which produced over 2,400 comments). The application window for new gTLDs opened on 12 January 2012 and the final Applicant Guidebook was published in June 2012.
Applying for a new gTLD is infinitely more complex than simply registering a domain name at the second level via a registrar. Successful applicants effectively control the domain name registry itself, a commitment that is not to be taken lightly. The ICANN application fee alone was $185,000, not including the additional significant costs of the necessary specialist and technical advice required in order for applicants to understand the rather complex application process and be guided through it.
A total of 1,930 applications for new gTLDs were submitted by the first round closing date of 30 May 2012. 230 of these were in competition with one or more exact matches (the most hotly contested string was .APP with 13 applications, followed by 11 for .HOME and .INC). On 23 October 2013, ICANN announced what it referred to as ‘an historic moment, not only for the new gTLD program, but for the internet as a whole’ because the first four new gTLD strings had completed the new gTLD process and been delegated into the root zone of the internet.
The first Sunrise Period (for the Arabic for ‘Web’ or ‘Network’) opened on 31 October 2013 and others are now following thick and fast. ICANN may soon reach its objective of delegating 20 new gTLDs per week. It is estimated that the first round of applications will result in around 1,400 new gTLDs, although ICANN has so far not given any detail on the timing of subsequent rounds. Evidently the launch of such a large number of new gTLDs in a relatively short space of time will rapidly result in a radical reshaping of the Internet and profoundly affect the way users search for and access corresponding web sites.
Community applications
Many different types of string were applied for, such as Internationalised Domain Names (commonly referred to as IDNs, for example .???? which is Russian for ‘Web site’), geographic names (for example .CYMRU or .LONDON), trade marks (for example .WALMART) or generic words (for example .CLOTHING). In fact such classifications are not mutually excusive and a string may fall into more than one category. Broadly speaking, all types of applications were subject to the same rules and procedures and it was not necessary for applicants to select a particular category. However, what applicants of all strings of whatever type had to do was to decide whether or not to designate their application as ‘community’. In the end only 84 applicants chose to do this, although those 84 applications covered all the different types of strings.
According to the Applicant Guidebook, a community-based gTLD is one that will be operated for the benefit of a clearly delineated community, as opposed to a standard gTLD which won’t (standard applications are thus effectively only defined by omission, and may or not propose eligibility criteria for domain name registration). Module 1 of the Applicant Guidebook gives a high-level overview of the application process and states that community applicants are expected to:
– demonstrate an ongoing relationship with a clearly delineated community;
– apply for a gTLD string strongly and specifically related to the community named in the application;
– propose dedicated registration and use policies for registrants in its proposed gTLD, including appropriate security verification procedures, commensurate with the community-based purpose it has named; and
– have their application endorsed in writing by one or more established institutions representing the relevant community.
To understand the implications of deciding to designate an application as community-based, it is necessary to have an overview of the different stages of the application procedure as a whole and how they fit together, as outlined by the five modules of the Applicant Guidebook. After the closing date, each of the 1,930 applications was subject to an Initial Evaluation (IE) by ICANN, the order of which was determined by a prioritisation draw. The IE process basically involved an examination of each application to check that it met the minimum standards in relation to various criteria, such as technical and operational capacity, financial capability, stability and the types of registry services proposed, all of which were scored in accordance with the detailed requirements set out in Module 2 of the Applicant Guidebook. Only those applications that passed IE were allowed to move on to the next stage of the program.
On 30 August 2013, ICANN announced that the IE phase had been concluded, representing a major milestone. Over 90% of the applications submitted (1,745) passed IE, and the rest were either withdrawn or went into the Extended Evaluation (EE) phase, which gives applicants the opportunity to rectify their failings. At the time of writing, EE is still ongoing for around 20 applications. The next stage for applicants after successful evaluation depends on factors particular to each application, and broadly speaking this relates to whether any objections have been made and/or whether any other identical or confusingly similar applications have been filed (known as string contention). The simplest applications, where no objections have been filed and there is no string contention, are allowed to proceed to the final transition to delegation phase, as set out in Module 5 of the Applicant Guidebook. This involves contractual negotiation of the registry services agreement with ICANN, and pre-delegation testing to ensure that applicants have the capacity to operate a new gTLD in a stable, secure manner.
The objection period for filing formal objections ran for nine months from 13 June 2012 to 13 March 2013, and in total over 260 objections were filed. The objection procedure is dealt with by Module 3 of the Applicant Guidebook, which sets out four different grounds for objection, including community objections. At the time of writing, many objections are still under consideration by the relevant dispute resolution providers. It is important to understand that it is only once all objections have been dealt with in relation to a particular string that such string is able to proceed to the string contention stage, dealt with by Module 4 of the Applicant Guidebook. However, this phase actually kicks in only if a string is in a contention set, meaning that one or more applicants have applied for a similar or identical string (similarity is based both on the outcome of ICANN’s string similarity review during the initial evaluation phase and also the results of any string confusion objections referred to above). Evidently the members of particular contention sets may be defined with certainty only after the evaluation and objection phases are complete.
In the event of contention that cannot be solved by self-resolution, if (and only if) one of the strings in question was self-designated as a community application upon filing of the application, that particular applicant may request Community Priority Evaluation (CPE). Only at this point will ICANN review the application to assess whether it may be given priority on a community basis. Applications are reviewed against four detailed criteria as follows:
– community establishment
– nexus between proposed string and community
– registration policies
– community endorsement.
Detailed definitions and guidelines for each requirement are contained within the Applicant Guidebook, and an applicant will require 14 out of 16 points to prevail in a CPE. The criteria are very stringent because a successful community application will effectively eliminate all standard applications in direct contention, no matter how well qualified they are. If the CPE fails and only standard applications remain then, if self resolution does not resolve the contention, an auction will be used as a method of last resort to decide who will be allocated the string in question. This will also be the case if two or more community applications in a contention set all pass CPE. Only four applicants have reached the stage of being able to request CPE, and this is currently still in progress for all of them.
It can therefore be seen that whether to self-designate as a community application at the beginning was a tricky strategic question for some applicants. Community designation allows an applicant to invoke CPE if this becomes necessary and thus increase the likelihood of winning the string but, in order to have any chance of success in a CPE, an application needs to fulfil the stringent criteria necessary, including with regard to the registration policies imposed on future registrants. To score highly the restrictions and corresponding enforcement mechanisms need to reflect the community-based purpose of the string, but this may have a huge impact on the number of potential registrations and consequent business model of the gTLD.
Whilst it is possible in theory to make a request to ICANN to change the registration policies of existing TLDs, such requests are not always successful. In any case, it would seem certain that ICANN would not take kindly to applicants trying to game the system by obtaining a coveted string as a result of CPE and then attempting to relax the registration rules once the registry is up and running. What’s more, it was not possible to make a community application and then change to a standard application once it became clear that there was no contention. The Applicant Guidebook emphasizes in Module 1 that community applications are expected to be a narrow category, meant for applications where there are unambiguous associations between the applicant, the community served and the applied-for new gTLD string.
The Applicant Guidebook is a long document containing detailed interconnecting instructions and guidelines and is impossible to summarise in a few neat paragraphs. However, the above highlights how the community concept is given key significance within the framework of the new gTLD program, reflecting the stated ICANN values of fairness and non-discrimination. In short, Module 3 deals with community objections and thus attempts to weed out applications that are harmful to communities, whilst Module 4 deals with any eventual contention and attempts to give priority to applications that favour communities. What is interesting from a legal point of view is that different definitions and standards apply to each Module. However, whilst Modules 3 and 4 are intricately linked, they are also clearly linear, and in fact the provisions of Module 4 and CPE will only fall to be considered in the event of contention.
Standards for community objections
The four types of objection may be briefly summarised as follows:
– String confusion: the string applied for is confusingly similar to an existing Top Level Domain or to another new gTLD applicant’s string. Complaints are administered by the International Centre for Dispute Resolution (ICDR).
– Legal rights: the string applied for violates the legal rights of the objector, namely registered or unregistered trade mark rights or the name or acronym of an intergovernmental organisation. Complaints are administered by WIPO.
– Limited public interest: the string applied for contradicts generally accepted legal norms of morality and public order recognised under principles of international law. Complaints are administered by the International Center of Expertise of the International Chamber of Commerce (ICC).
– Community: there is substantial opposition to the string applied for from a significant portion of the community that the gTLD string targets. Complaints are administered by the ICC.
A total of 104 community objections were filed against applications for new gTLD strings. A point that is often misunderstood is that community objections may be filed against any string, not just strings that have been designated as community applications. So far there have only been around 20 decisions issued by the ICC, based in Paris, although a significant number are expected to be published before the end of the year. The Attachment to Module 3 of the Applicant Guidebook sets down the New gTLD Dispute Resolution Procedure covering all four different types of objections and in all cases the particular standards for each type of objection are set down by the Applicant Guidebook.
All objectors must first satisfy standing requirements in order to have their objections considered and these vary depending on the type of objection. As far as community objections are concerned, the Applicant Guidebook provides (at Module 3, para. 3.2.2.4) as follows:
‘Established institutions associated with clearly delineated communities are eligible to file a community objection. The community named by the objector must be a community strongly associated with the applied for gTLD string in the application that is the subject of the objection. To qualify for standing for a community objection, the objector must prove both of the following:
It is an established institution – Factors that may be considered in making this determination include, but are not limited to:
• Level of global recognition of the institution;
• Length of time the institution has been in existence; and
• Public historical evidence of its existence, such as the presence of a formal charter or national or international registration, or validation by a government, inter-governmental organization, or treaty. The institution must not have been established solely in conjunction with the gTLD application process.
It has an ongoing relationship with a clearly delineated community – Factors that may be considered in making this determination include, but are not limited to:
• The presence of mechanisms for participation in activities, membership, and leadership;
• Institutional purpose related to the benefit of the associated community;
• Performance of regular activities that benefit the associated community; and
• The level of formal boundaries around the community.
The panel will perform a balancing of the factors listed above, as well as other relevant information, in making its determination. It is not expected that an objector must demonstrate satisfaction of each and every factor considered in order to satisfy the standing requirements.‘
The Applicant Guidebook also sets down the dispute resolution principles or standards that need to be fulfilled for each type of objection. For community objections, the Applicant Guidebook sets out four tests that must be satisfied to enable Panels to determine whether there is substantial opposition from a significant portion of the community to which the string may be targeted, as follows (at Module 3, para. 3.5.4):
‘For an objection to be successful, the objector must prove that:
• The community invoked by the objector is a clearly delineated community; and
• Community opposition to the application is substantial; and
• There is a strong association between the community invoked and the applied-for gTLD string;
and
• The application creates a likelihood of material detriment to the rights or legitimate interests of a significant portion of the community to which the string may be explicitly or implicitly targeted.’
The Applicant Guidebook goes on to elaborate on each of these four tests in some detail.
It can therefore be seen that the Applicant Guidebook sets a very high bar for successful community objections, and so it is not surprising that the majority have so far failed, allowing the applicant to prevail. This does not necessarily mean that the successful applicant will be awarded the string in question though as it must first prevail in any other objections filed against it and then triumph in the event of contention (either via CPE, if possible, or via an auction).
It should also be noted that a successful objection does not necessarily mean that the string in question will not be delegated, because any rival applicants for the same string who have not had objections filed against them (or who have won) may continue the process. However, a losing applicant must withdraw from the process ($37,000 of the original $185,000 application fee will be refunded, although this is scant consolation for most applicants). This is true for all types of objections apart from string confusion objections where failure for an applicant simply means that the affected string will be placed in a contention set with the objector’s applied-for string and possibly others (except if the successful objector was an existing Top Level Domain, in which case the application will fail).
Decisions following community objections
Panels making early decisions in any new procedure are in virgin territory and thus faced with difficult choices. Most have therefore chosen to play safe and methodically go through the above requirements and standards in the order that they are set out in the Applicant Guidebook. However, a careful reading of them reveals that this is slightly problematic as they all require the presence of a clearly delineated community, as per the first limb of the four step test, but the standing requirement falls to be considered first. This makes for a certain amount of repetition in some cases (see Case No. EXP/429/ICANN/46 (.REISEN), para 50 for a good explanation of these issues ).
A further difficulty is that, whilst it is primordial for objectors to prove the existence of a clearly delineated community under Module 3, Module 3 itself does not elaborate on the notion of ‘community’, merely on those factors that may assist in deciding whether such community is ‘clearly delineated’ or not. To this end, certain Panels have drawn guidance from the wording of the final report of the GNSO in 2007, which contains the following definition of ‘community’ in relation to community objections:
‘community should be interpreted broadly and will include, for example, an economic sector, a cultural community, or a linguistic community. It may be a closely related community which believes it is impacted.’
To add to the confusion, Module 4 of the Applicant Guidebook does contain a definition of ‘community’, but this is in relation to community establishment, the first of the four criteria for CPE:
‘Usage of the expression “community” has evolved considerably from its Latin origin – “communitas” meaning “fellowship” – while still implying more of cohesion than a mere commonality of interest. Notably, as “community” is used throughout the application, there should be: (a) an awareness and recognition of a community among its members; (b) some understanding of the community’s existence prior to September 2007 (when the new gTLD policy recommendations were completed); and (c) extended tenure or longevity—non-transience—into the future.’
As highlighted by certain Panels (see further below in relation to the objections by ILGA to .GAY for example), CPE under Module 4 is intended to be a separate procedure and the standards set out therein may only be invoked once all objections under Module 3 have been dealt with. Thus strictly speaking the definition in Module 4 does not apply to objections, although this makes matters rather more complicated than they need to be.
Of those community objections that have succeeded, it is significant that a number relate to strings that could be said to belong to highly regulated areas, for example .ARCHITECT, .BANK, and .MEDICAL. This does not mean that these types of strings are not permitted (indeed it is likely that one of the two applications for .LAWYER will eventually succeed because no objections were filed against either of them), but simply that the objectors in these cases were able to fulfil not only the standing requirements but also the stringent four-step test, including likelihood of material detriment. The applicants for .ARCHITECT and .BANK were proposing to implement unrestricted registration policies (which was also a subject of concern for various governing bodies and national governments), and did not previously have strong ties to the industry/profession (the applicant for .BANK effectively admitted that it had no real familiarity with the highly complex world of national and international banking regulation).
It should be underlined that many objections have been made by those communities backing rival applicants for the same or a similar string and so could be said to be strategic in nature, and this was true for both .ARCHITECT and .BANK. However the .MEDICAL objection was slightly different in that it was made by the Independent Objector (IO), Professor Alain Pellet, who was appointed by ICANN to consider the applied-for strings and make either community or limited public interest objections to ‘highly objectionable’ strings in the best interests of the public who use the global Internet. After consideration of the applied-for strings, the IO made a total of 24 objections, 13 of which were community objections, including .MEDICAL. The IO is exempt from the requirement to prove standing, but must still pass each requirement of the four step test in order to succeed.
So far only one unfortunate trade mark owner has been knocked out by a successful community objection: Ralph Lauren Corporation (RLC), the applicant for .POLO. However it is envisaged that such cases will be relatively rare as most brand names do not call to mind a separate community. The Panel found in favour of the objector, the United States Polo Association (USPA), on all four grounds, but this does not mean the string will be awarded to USPA as it did not make an application. However it will remain undelegated, and this will enable the USPA to apply in the next round if it so desires.
Turning to the unsuccessful objections, the majority of objectors have nevertheless satisfied the basic standing requirements, although there have been some that have fallen at the first hurdle. For example, four applications were made for the .GAY string, and four community objections were filed, three by The International Lesbian Gay Bisexual Trans and Intersex Association (ILGA) and one by Metroplex Republicans of Dallas (Metroplex), a political organization in Texas that raises public awareness of gay conservatism. Whilst ILGA satisfied the standing requirements (although its objections ultimately went on to fail), Metroplex did not. In the Panel’s view, Metroplex was associated with the conservative segment of the gay community (as opposed to the gay community as a whole) which was not a clearly defined community in and of itself. The Panel in Metroplex Republicans of Dallas v dotgay llc, Case No. EXP/446/ICANN/63 explained this as follows (at para 13):
‘That some LGBTQ people hold conservative political views and vote for conservative candidates may bring them into a statistical category, but does not make them connect, gather, interact, or do anything else together that would constitute a community, or, that would make them publicly visible as one. That people hold a political view or vote for a political candidate also does not mean that they do so consistently and stably over time. People change their political views, and have different views on different issues; and whether a political view is to be characterized as conservative, liberal, or something else is often debatable.‘
The Panel also went on to find that, even if Metroplex was held to represent the gay community as a whole, it would still lack standing because it did not have an ongoing relationship with such community and so did not fulfill the relevant factors listed in the Applicant Guidebook.
Most objectors were found to have standing though, and thus in turn to pass limbs one and three of the four-step test relating to a clearly delineated community and a strong association between the applied-for gTLD string and the community represented by the objector, as these issues are tightly bound together. However, so far all unsuccessful objections (where the Panel has actually needed to consider the four step test) have failed to pass limb four and prove that the application creates a likelihood of material detriment to the rights or legitimate interests of a significant portion of the community to which the string may be explicitly or implicitly targeted.
At this point it is useful to recall the fact that Modules 3 and 4 of the Applicant Guidebook are clearly separate, and at the objection stage Panels are required to focus on reasons why the applied-for string should not be delegated (if any), as opposed to why any competing applications should be given priority. Indeed, the Applicant Guidebook specifically states at Module 3, para 3.5.4 that ‘An allegation of detriment that consists only of the applicant being delegated the string instead of the objector will not be sufficient for a finding of material detriment‘. This is illustrated particularly well by the three decisions relating to ILGA’s objections to .GAY, which all failed because ILGA (a competing applicant) failed to prove that the other three applications would create a likelihood of material detriment to the gay community. The Panel in The International Lesbian Gay Bisexual Trans and Intersex Association (Belgium) v Top Level Design, LLC (USA), Case No. EXP/392/ICANN/9 commented (at para 30):
‘The interplay between the Community Objection Procedure of Module 3.2.2.4 of the Guidebook and the Community Priority Evaluation Procedure of Module 4.2.2 of the Guidebook is intricate. The contention between two applicants, one a community-based applicant, the other a standard applicant, and the decision on whether the community-based applicant will serve the community involved well enough to win against the standard applicant, belong into the Community Priority Evaluation Procedure of Module 4.2.2 of the Guidebook. If the community involved could exclude the other applicant by using the Community Objection Procedure of Module 3.2.2.4 of the Guidebook, presenting an established institution as an objector, and presenting arguments that were not sufficient to win in the Community Priority Evaluation Procedure of Module 4.2.2 of the Guidebook, this procedure would be voided and kept from serving the purpose for which it is created.‘
A number of decisions have also failed on the grounds of limb two as well as limb four, namely failure to prove that community opposition is substantial, for example those relating to .HALAL and .ISLAM (Case Nos. EXP/427/ICANN/44 and EXP/430/ICANN/47). In many incidences though Panels have found substantial opposition but have nevertheless gone on to decide that the objector has failed to prove material detriment (as in the case of .PERSIANGULF, for example: see Case No. EXP/423/ICANN/40).
Conclusion
Despite its length, the above only scrapes the surface of the complexities of the Applicant Guidebook and in particular how it defines and deals with communities. ICANN has been much criticised in relation to its new gTLD program, but trying to anticipate and deal with all eventualities in relation to situations that have never been seen before is certainly not easy. Hopefully lessons will be learned from this first round of applications that will clarify and improve the process going forward. As far as communities are concerned, it remains to be seen whether the community objection process will provide an efficient mechanism to weed out applications that will be harmful to communities and whether CPE will ensure that communities are given the boost they need to secure the new gTLD they deserve.
Jane Seager is of Counsel at Hogan Lovells International LLP in the Intellectual Property, Media and Technology Group in Paris. She specializes in Internet related intellectual property, and in particular the protection of brands online.
